Anesthetic Management

All children should be monitored with standard monitoring that includes a pulse oximeter, a capnograph for end-tidal carbon dioxide (CO₂) monitor, precordial stethoscope, electrocardiogram, thermometer, and automated blood pressure cuff. Often, electromyography is used to avoid injury to the facial nerve. In these circumstances, neuromuscular blockade is avoided. Standard no oral intake (NPO) guidelines should be followed in nonemergent situations.

For mastoidectomy and tympanoplasty, the patient is positioned supine, with the head laterally rotated away from the affected side. It is important to carefully rotate the neck of a patient under general anesthesia because of the risk for atlantoaxial rotational subluxation (Brisson et al., 2000).

Induction of anesthesia can entail inhalational or intravenous induction. The anesthesiologist should consider endotracheal intubation without muscle relaxation or with a short-acting muscle relaxant so the facial nerve can be monitored during surgery. The use of nitrous oxide may be contraindicated during and after placement of the tympanic graft, because nitrous oxide accumulates in a closed gas space and increases the ambient pressure; some surgeons believe this tends to lift the graft away from its new site (Koivunen et al., 1996; Doyle and Banks, 2003).

The most common postoperative complication requiring unscheduled admission to the hospital for children who have undergone tympanomastoidectomy is postoperative nausea and vomiting (PONV) (Megerian et al., 2000). PONV has been attributed to surgical stimulation of the vestibular labyrinth, anesthetic techniques, or a combination of the two (Jellish et al., 1995; Dornhoffer and Manning, 2000). Propofol has antiemetic properties (Ved et al., 1996; Erb et al., 2002b). It appears that propofol used for induction and maintenance of anesthesia is superior to isoflurane or sevoflurane in reducing PONV after middle ear surgery (Jellish et al., 1995, 1999; Moore et al., 2003). The use of intravenous dexamethasone during surgery appears to decrease PONV in patients after undergoing tympanomastoid surgery (Liu et al., 2001).

Obstructive Sleep Apnea Syndrome

Obstructive sleep apnea syndrome (OSAS) is a disorder of breathing during sleep characterized by prolonged partial upper airway obstruction (obstructive hypopnea) or intermittent complete obstruction (OSA) with or without snoring and associated with moderate to severe oxygen desaturation that disrupts normal sleep-time breathing and normal sleep patterns (American Thoracic Society, 1996). The most common cause of OSAS among children is upper airway narrowing with adenotonsillar hypertrophy. OSAS also occurs in infants and children with upper airway narrowing resulting from craniofacial anomalies, and in those with neuromuscular diseases, including cerebral palsy and muscular dystrophy (Marcus, 2001; Schwengel et al., 2009) (Box 24-1). See related video online at www.expertconsult.com.

In recent years, the epidemic increase in the prevalence of obesity during childhood seems to be contributing to substantial changes in the cross-sectional demographic and anthropometric characteristics of the children being referred for evaluation of OSAS. Although less than 15% of all symptomatic habitually snoring children were obese (i.e., >95th percentile for age and gender) in the early 1990s, more than 50% fulfilled the criteria for obesity among all referrals to a Kentucky sleep center by the mid 2000s (Gozal et al., 2006; Schwengel et al., 2009).

Adenotonsillar hypertrophy is the most common cause of OSAS in children. Many published papers, primarily case reports and case series, support the idea that tonsillectomy with or without adenoidectomy is often the cure for OSAS (Schechter et al., 2002; Garetz, 2008).

The peak prevalence of OSAS in children occurs between 2 and 8 years, which is the age when the tonsils and adenoids are large in relation to the child’s upper airways. The site of collapse is most commonly at the level of the adenoids or velopharynx (Isono et al., 1998; Isono, 2006). Although OSAS is associated with adenotonsillar hypertrophy, there must be other neuromuscular factors involved. These patients usually do not obstruct while awake, implying that sleep induces another dimension to OSAS. Some otherwise normal children with OSAS but with smaller tonsils and adenoids are cured by adenotonsillectomy, whereas others with larger tonsils and adenoids are not (Marcus, 2001). The patency of pharyngeal airway is maintained by tonic and phasic contractions of the upper airway dilator muscles, such as the genioglossus, geniohyoid, and velopalatine muscles (Isono, 2006). Compared with other inspiratory muscles (i.e., diaphragm and intercostal muscles), these upper airway muscles are preferentially depressed with sleep, sedatives, and general anesthetics (Ochiai et al., 1989, 1992) (see Chapter 3, Respiratory Physiology in Infants and Children). It is hypothesized that children with OSAS may have abnormal centrally mediated activation of their airway muscles, leading to a more collapsible upper airway (Marcus et al., 1994).

Fatty deposits of the upper airway structures, and subcutaneous fat deposits in the anterior neck region are the direct causes of airway obstruction in obese children (mostly adolescents), as well as in obese adults, with OSAS. In addition, increased adipose deposits in the chest wall and abdomen can reduce chest wall compliance as well as cephalad displacement of the diaphragm, causing inefficient ventilation by mass loading and decreased functional residual capacity of the lung (Dayyat et al., 2007; Schwengel et al., 2009). Obese patients have symptoms that differ from those in children with OSAS caused by adenotonsillar hypertrophy. Obese patients tend to have a significantly greater incidence of daytime sleepiness, insulin resistance, and systemic as well as pulmonary hypertension than children with adenotonsillar hypertrophy (Tauman et al., 2004). Furthermore, Tauman and coworkers recently observed a high failure rate for obese children with OSAS undergoing adentotonsillectomy.

Dayyat and colleagues (2007) differentiated between children with OSAS into two types. The type 1 child with OSAS is not obese but has disordered breathing secondary to marked lymphadenoid hypertrophy. The type 2 patient is obese with minimal lymphadenoid hyperplasia. Table 24-1 compares the common symptoms experienced by the two types of patients (Dayyat et al., 2007).

TABLE 24-1 Clinical Presentation of Pediatric Obstructive Sleep Apnea Types I and II

−, Absent; + infrequent to ++++ very frequent.

From Dayyat E, Kheirandish-Gozal L, Gozal D: Childhood obstructive sleep apnea: one or two distinct disease entities? Sleep Med Clin 2:433, 2007.

One of the most interesting areas of childhood OSAS research has been on the effects that OSAS has on systemic inflammation. Children with OSAS are at risk for developing inflammatory responses that may lead to endothelial dysfunction and atherogenesis (Gozal et al., 2008). School-aged children with OSAS have a higher serum C-reactive protein level than controls, and the elevation correlates with the severity of OSAS, and levels decrease after adenotonsillectomy (Gozal et al., 2007). Serum proteomic patterns in children with OSAS differ significantly from those in children who have primary snoring but do not have OSAS (Shah et al., 2006). In the Shah group’s study, proteomic profiling of serum samples in children with OSAS revealed differential expression of circulating proteins. Proteomics may play a future role in diagnosing OSAS in the snoring child.

Symptoms of OSAS include nocturnal snoring, breathing pauses, gasping, use of accessory muscles of respiration, enuresis, and excessive sweating (Messner, 2003). In addition, children with OSAS have a host of sequelae, which are usually reversible after adenotonsillectomy but can lead to perioperative complications during and after surgery. Children with OSAS may present with failure to thrive, although the cause of this growth failure is unclear (Sterni and Tunkel, 2003). Many of these children experience a growth spurt after surgery. Children with OSAS may have neurocognitive deficits such as poor learning, behavioral problems, and lower grades in school than non-OSAS children. Adenotonsillectomy improves functioning in these children (Marcus, 2001). Recent studies demonstrate that despite negative polysomnographic findings, the benefits of adenotonsillectomy are still observed in children with symptomatic airway obstruction, including improvements in behavior and quality of life (Leong and Davis 2007).

The diagnosis of OSAS is based on a thorough history and physical examination along with appropriate sleep studies, including polysomnography (Table 24-2 and Boxes 24-2 and 24-3). Snoring, increased respiratory efforts, periodic obstructive apnea, and oxygen desaturation while sleeping are the universal features of OSAS, which must be differentiated from the benign condition referred to as primary snoring.

TABLE 24-2 Polysomnographic Characteristics of Obstructive Sleep Apnea in Children and Adults

From Karlson KH Jr: What’s new in pediatric obstructive sleep apnea? Clin Pulm Med 15:226, 2008.

Polysomnography

Pediatric polysomnography includes measurement of end-tidal CO₂, electroencephalogram, chin EMG, chest wall movement, airflow through the nose and mouth, leg movement, and oxygen saturation via pulse oximetry (Box 24-4). After all events are reviewed and scored, several indices can be calculated. These include the apnea index and the apnea-hypopnea index, which are calculated for the entire sleep period, for both rapid eye movement sleep and non–rapid eye movement sleep. The term index refers to the number of events divided by the number of hours of sleep (Table 24-3). This calculation allows the comparison of polysomnograms of varying time lengths. The apnea index is determined using only apneas, whereas the apnea-hypopnea index includes both apneas and hypopneas (Wagner and Torrez, 2007).

TABLE 24-3 Respiratory Events that Can Be Seen During Polysomnography

It is helpful to divide the scores into severity-based categories (Table 24-4). OSAS severity can be used to determine the appropriate surgical center where the surgery should be performed and whether the child requires overnight observation (Box 24-5).

Although definitive diagnosis of OSAS is made by a positive polysomnogram, many children are not tested because the polysomnography laboratories specializing in children are few, and the tests are expensive and require an overnight stay (Sterni and Tunkel, 2003). Most children are diagnosed by recommendations set forth by the section of Pediatric Pulmonology Subcommittee on Obstructive Sleep Apnea, sans Polysomnography. This committee’s recommendations for otherwise healthy children older than 1 year are as follows (Schechter, 2002):

Severe, untreated OSAS may lead to pulmonary hypertension and cor pulmonale caused by nocturnal hypoxia and hypercarbia, resulting in compensatory changes in the pulmonary vasculature. Pulmonary vascular resistance increases, causing increased right ventricular strain. Severe cases may progress to pulmonary hypertension, arrhythmias, and cor pulmonale, which are reversible by performing early adenotonsillectomy (Miman et al., 2000). Children with OSAS also tend to have higher diastolic blood pressures. The cardiovascular changes appear to be the result of an increase in sympathetic tone that results from obstructive respiratory events (Marcus et al., 1998). Fortunately, few children develop clinically significant heart failure. It is prudent to pursue an aggressive cardiac evaluation if the child has a loud second heart sound, exercise intolerance, or diastolic hypertension. Because children are diagnosed earlier today, significant heart disease is rarely an issue (Marcus, 2001).

Preoperative Preparation

A careful review of the history, laboratory data, and physical examination results are essential for the optimal outcome of adenotonsillectomy. Careful evaluation of coagulation status is important before performing this procedure. If there is a history of easy bruising, frequent epistaxis, or positive family history, the prothrombin time, partial thromboplastin time, and platelet count should be obtained to rule out coagulopathies. In many pediatric institutions, a coagulation profile is obtained routinely for patients scheduled for adenotonsillectomy. However, in a prospective study of hemostatic assessment of patients before tonsillectomy, routine measurements of a coagulation profile were not useful predictors of postoperative bleeding (Close et al., 1994).

Although relatively mild, von Willebrand’s disease (i.e., reduced factor VIII, decreased platelet adhesiveness because of deficient von Willebrand factor) is the most common coagulopathy seen among patients scheduled for adenotonsillectomy. Children with von Willebrand’s disease or mild hemophilia A are treated preoperatively with desmopressin (1-desamino-8- d-arginine vasopressin [DDAVP]) (Prinsley et al., 1993). DDAVP is a synthetic analogue of vasopressin, which, in addition to its antidiuretic effect, stimulates endothelial cells and releases stored factor VIII and von Willebrand factor (Mannucci, 1988). The intravenous dosage of DDAVP is 0.3 mcg/kg, given over 20 minutes before anesthetic induction.

Patients who receive DDAVP are at increased risk for hyponatremia because of the retention of free water that occurs with DDAVP. Anesthesia care providers need to remember that intravenous fluids for these patients should be isotonic (preferably normal saline) and should be administered at one-half to two-thirds maintenance dosage. Patients who require intravenous fluids postoperatively should have their sodium monitored to assess for the development of hyponatremia.

The examiner should elicit a history of drug ingestion, especially acetylsalicylic acid. If the patient was given such drugs recently, surgery should be postponed, because these drugs cause platelet dysfunction for as long as 10 days and may cause excessive bleeding intraoperatively and postoperatively (Davies and Steward, 1977; Paradise, 2003).

During the preoperative visit, the patency of oral and nasal air passages is carefully examined. The patient’s mouth should be inspected for the degree of tonsillar hypertrophy or inflammation. The examiner should also have the child breathe with the mouth closed to evaluate the degree of nasal airway obstruction and estimate adenoidal hypertrophy. It is important to inspect the teeth routinely, because tonsillectomy is often performed on children who are losing their primary dentition. Any teeth missing preoperatively should be noted carefully. A tooth that is loose should be pointed out to the parent and child, with the explanation that it may be necessary to remove it while the child is anesthetized (it must then be saved). It is also possible for the surgeon to dislodge or chip teeth in the application of a mouth gag or during other intraoperative manipulations.

Anesthetic Management

All children should be monitored with a pulse oximeter, end-tidal CO₂ precordial stethoscope, electrocardiogram, thermometer, and automated blood pressure cuff. Oxygen saturation is determined in room air before the induction of anesthesia to establish the baseline value. If neuromuscular blockade is required, a nerve stimulator should be used to track the depth of paralysis. Standard NPO guidelines should be followed in nonemergent situations.

Children who are anxious preoperatively may receive sedation with oral midazolam. In older children scheduled for intravenous induction, EMLA cream (i.e., eutectic mixture of local anesthetics, 2.5% lidocaine and 2.5% prilocaine) is applied to the dorsum of both hands and sealed with plastic adhesives at least 1 hour before intravenous catheter insertion (see Chapter 9, Preoperative Preparation, and Chapter 13, Induction, Maintenance, and Recovery).

Adenoidectomy is a relatively short procedure (15 to 45 minutes), and many centers perform anesthesia without neuromuscular blockade and with less opioid than suggested in the following descriptions. Anesthesia is induced most commonly with oxygen, nitrous oxide, and sevoflurane. As with those undergoing myringotomies, children with adenotonsillar hypertrophy often have partial or complete nasal airway obstruction. The mouth must be kept open during the induction of anesthesia until the gag reflex is abolished and the oral airway is inserted. If necessary, moderate CPAP (10 to 15 cm H₂O), with jaw thrust during induction (before the patient is deep enough for oral airway insertion) helps to prevent the pharyngeal airway collapse that results from the relaxation of upper airway muscle tone (Reber et al., 2001; Bruppacher et al., 2003).

After the intravenous route is established, atropine or glycopyrrolate may be given for its anticholinergic effects. The patient’s trachea is intubated with a preformed oral (RAE) tube under deep inhalational anesthesia supplemented with a topical lidocaine spray to the vocal cords under a direct vision, an intravenous bolus of propofol, or an intermediate-acting, nondepolarizing muscle relaxant (e.g., cisatracurium, vecuronium). In older children, anesthesia may be induced intravenously with propofol (2 to 3 mg/kg) mixed with lidocaine (1 to 2 mg/mL of propofol) to reduce pain at the injection site, with or without anticholinergic agent or a muscle relaxant. Inhalation anesthetic or propofol infusion is continued thereafter with spontaneous or controlled ventilation. A cuffed endotracheal tube is recommended to reduce the chance of aspiration of blood and secretions and to reduce gas leaks around the tube. A cuffed endotracheal tube, 0.5 to 1.0 mm (inner diameter) smaller than the age-appropriate size, should be chosen to accommodate the passage of the cuff through the subglottis (Khine et al., 1997; James, 2001; Fine and Borland, 2004). The endotracheal tube is immobilized with adhesive tape over the middle of the lower lip (Fig. 24-4). The breath sounds on both sides of the chest should be auscultated carefully to avoid endobronchial intubation. The endotracheal tube is then held in place by the groove-bladed tongue depressor that is a part of the Ring adaptation of the Brown-Davis mouth gag (Fig. 24-5).

FIGURE 24-4 For tonsillectomy or adenoidectomy, an oral Ring, Adaire, Ellwyn (RAE) endotracheal tube is fixed over the middle of the lower lip.

FIGURE 24-5 Patient in position for tonsillectomy. A Brown-Davis mouth gag holds the mouth open. An oral Ring, Adaire, Ellwyn (RAE) endotracheal tube is fixed to the middle of the lower lip and is held in the groove of the tongue blade.

Anesthesia is maintained with supplemental opioids to reduce the requirement of the anesthesia maintenance agent and to provide postoperative analgesia. The physician may start with a loading dose of 1 to 2 mcg/kg of fentanyl, 50 to 100 mcg/kg of morphine, or 10 to 20 mcg/kg of hydromorphone given intravenously, with additional doses administered as needed. Children with moderate and severe OSAS should receive lower dosages of opioids to minimize the risk for prolonged apnea during emergence and postoperative upper airway obstruction. Brown and coworkers (2004) have shown that young age and low preoperative oxygen saturation were associated with lower postoperative morphine requirement. It appears that children with a minimal oxygen saturation of 85% or less during preoperative sleep required half the postoperative analgesic dosage required by children whose saturations were 85% or greater (Brown et al., 2006).

The use of NSAIDs during or after tonsillectomy is controversial. A meta-analysis suggested that there is an increased risk for bleeding and increased return to the operating room when NSAIDs such as ketorolac are used (Marret, 2003). Criticism of this meta-analysis focuses on the small number of patients evaluated, the varying types and dosages of NSAID, and the varying surgical techniques used for the tonsillectomy. A Cochrane database review evaluated the role of NSAIDs in posttonsillectomy bleeding and found no significant correlation between NSAID use and increased risk for posttonsillectomy bleeding (Cardwell et al., 2005). Practice among pediatric anesthesiologists varies. Some centers do not use NSAIDs during adenoidectomy or tonsillectomy, whereas others use them routinely (Allford, 2009).

There is evidence that high dosages of dexamethasone (up to 1 mg/kg, 25 mg maximum) reduce postoperative swelling and pain and decrease the incidence of PONV without apparent adverse effects attributable to dexamethasone (Pappas et al., 1998). Children receiving dexamethasone are more likely to advance to a soft-solid diet on the first postoperative day (Steward et al., 2003).

Bleeding during tonsil and adenoid surgery is seldom excessive, but massive hemorrhage has occurred with tearing of the carotid vessels (Smith, 1972, 1980). Homeostasis is usually obtained using electrocautery. There have been several reports of fires due to electrocautery-induced ignition of the endotracheal tube or packing during tonsillectomy. Fires are caused by combustible material in an oxygen-rich environment (Mattucci and Militana, 2003); management of airway fires is discussed later.

Children with OSAS tend to emerge from anesthesia more slowly than children without OSAS. This may be explained by their deficit in sleep arousal mechanisms. They seem to have elevated sleep arousal mechanisms in response to hypercarbia and increased upper airway obstruction (Marcus et al., 1998, 1999). Other subtle disturbances of sleep architecture may also be present (Bandla et al., 1999).

Techniques of Tonsillectomy and Adenoidectomy

A myriad of surgical techniques are used to perform tonsillectomy and adenoidectomy, and they use a mouth gag to give oropharyngeal exposure. The surgeon should use caution when placing and removing the mouth gag to avoid displacement of the endotracheal tube.

The adenoidectomy is performed either with a mechanical or a thermal technique (Hesham, 2009). Mechanical removal of the adenoids may be performed using curettes, adenotomes, or a micro-debrider. These mechanical techniques require nasopharyngeal packing to prevent bleeding, and the surgeon must remember to remove the packing at the end of the procedure. Thermal removal of adenoid tissue may employ electrosurgical cautery, radiofrequency ablation, or lasers, and it usually results in little bleeding.

Tonsillectomy can also be performed via mechanical or thermal techniques. Intracapsular removal of tonsil tissue is a newer approach that preserves the integrity of the tonsil capsule, and will be discussed later.

Mechanical removal of the tonsil has many technical variations. In general, an incision is made in the anterior pillar mucosa to expose the tonsil capsule, and using sharp or blunt dissection, the tonsil and capsule are removed together. Alternatively, a guillotine, or snare, is occasionally used to remove the tonsil without making an incision. This results in exposure of the underlying pharyngeal musculature with the perforating vascular supply to the tonsil, and hemostasis must be established. At the turn of the previous century, snare tonsillectomy with ether anesthesia was the standard, and hemostasis was achieved after prolonged observation of the surgical field to ensure passive wound clotting. This was less than ideal, and the mortality rate was unacceptably high. Today, active hemorrhage control is performed after removal of the tonsils. Some surgeons use dissolvable suture to ligate the bleeding points, but this is becoming less common. More often, electrosurgical cautery is used for hemostasis. This technique completes an electrical arc between the patient and the surgical instrument and has the potential to ignite flammable gases and combustible materials (such as the sponges in the nasopharynx used for tamponade of the adenoid bleeding). Case reports of airway fires have been reported with the use of electrosurgical cautery in tonsillectomy (Mattucci and Militana, 2003). It is suggested that the inspired oxygen levels be maintained as they are for the use of lasers in the airway (see Laser Surgery of the Larynx, later).

Thermal removal of the tonsil uses the electrosurgical unit or a radiofrequency unit for tonsil dissection. The most common radiofrequency unit in use is Coblation (ArthroCare, Austin, TX). Thermal removal of the tonsils and adenoids generally results in less blood loss, but it may be accompanied by slightly more postoperative pain than other techniques. A cuffed endotracheal tube, using moist throat packing and inspired oxygen levels of less than 40%, will minimize the risk for an airway fire (Johnson et al., 2002).

Intracapsular tonsillectomy is a newer technique that uses either the micro-debrider or the radiofrequency ablator to remove the tonsil tissue while preserving the tonsil capsule (Koltai et al., 2002). In effect, this may be considered a selective removal of the tonsillar mass, which is believed to be the most common cause of OSAS. Preservation of the capsule protects the underlying pharyngeal musculature, resulting in less postoperative pain and hemorrhage. Because the vascular supply is interrupted at a more terminal point, the muscular bed can contract around the vessels, providing better hemostasis. Initially, this technique was recommended only for those patients with OSAS. Recurrent infection is considered a contraindication, because this technique leaves behind residual tonsil tissue that may become reinfected. This technique is especially appropriate in those OSAS children 3 years of age and younger because of the lower postoperative pain scores and thus less opioid need (Solares et al., 2005; Colen et al., 2008). In general, the only risk of an intracapsular tonsillectomy is the rare (<1 in 1500) need to return for a complete tonsillectomy because of tissue regrowth leading to recurrent infectious tonsillitis.

At the conclusion of surgery, the surgeon usually suctions the pharynx and larynx under direct vision to prevent bleeding caused by agitation of raw mucosal surfaces by a suction catheter. Suctioning of the stomach contents may also remove swallowed secretions, blood, or residual preoperative sedative medications. Removal of these preoperative medications may decrease excess sedation in the postoperative period. Some surgeons elect to infiltrate 0.25% bupivacaine or other local anesthetic medications into the tonsillar fossa to aid in postoperative pain relief (Naja et al., 2005), whereas others feel that this analgesic technique may lead to premature discharge of patients who will then experience rebound pain when the local anesthetic wears off. This practice has occasionally been associated with airway obstruction after extubation.

After tonsillectomy, the anesthesia provider should always use a tonsil suction tip and should never blindly suction the nasopharynx or oropharynx after adenoidectomy or tonsillectomy. Suctioning should be directed toward the midline to avoid the lateral tonsil fossa. A soft nasopharyngeal airway may be inserted to facilitate airway patency for extubation. If adenoid hemostasis is adequate prior to insertion of the nasopharyngeal airway, usually no additional bleeding is encountered. Toward the completion of surgery, before a child regains active reflexes, the anesthesiologist should auscultate both sides of the chest to rule out the presence of aspirated blood or secretions and, under direct laryngoscopic visualization, examine the mouth and pharynx for blood and other debris that could cause airway irritation after extubation.

Laryngospasm may occur when the patient is extubated after tonsillectomy. Methods for avoiding this problem include extubating the trachea while the child is deeply anesthetized (not recommended for children with OSAS) or almost completely awake (see Chapter 13, Induction, Maintenance, and Recovery). Although intravenous lidocaine has not consistently proved helpful in preventing laryngospasm, application of local anesthetic before intubation or manipulation of the airway can reduce the incidence of laryngospasm during intubation and after extubation (Leicht et al., 1985; McCulloch et al., 1992; Landsman, 1997). It appears that lidocaine deposited locally in the laryngeal area suppresses laryngeal mucosa neuroreceptor transmission. However, lidocaine’s duration of action at laryngeal receptor sites is only 30 minutes, and its administration before intubation may not protect against laryngospasm at the time of extubation (Warner, 1996).

Before extubation, the lungs should be well expanded with an oxygen-air mixture at sustained high positive pressure (35 to 40 cm H₂O) several times (the vital capacity maneuver), to reopen intrathoracic airways and reverse atelectasis, which frequently develops during anesthesia and surgery; the patient should then be extubated under positive pressure (Benoit et al., 2002; Tusman et al., 2003) to prevent postoperative oxygen desaturation. Positive airway pressure at the moment of extubation causes a coughing motion, which helps to expel secretions around the vocal cords. Positive extending pressure on the upper airway walls also attenuates the excitation of the superior laryngeal nerve and may diminish the risk for laryngospasm (Suzuki and Sasaki, 1977; Sasaki, 1979) (see Chapter 3, Respiratory Physiology in Infants and Children, and Chapter 13, Induction, Maintenance, and Recovery).

Postoperative Management

Children with hypertrophic tonsils and adenoids tend to have increased airway obstruction in the immediate postoperative period. The presence of blood and secretions in the pharynx and larynx may provoke upper airway reflexes, leading to laryngospasm. These patients tend to become hypoxemic more often and perhaps more severely during the first several hours after surgery than patients undergoing procedures not involving the upper airways (Motoyama and Glazener, 1986). However, they seem to maintain their preoperative levels of oxygen saturation in room air thereafter, as determined with pulse oximetry (Motoyama and Borland, unpublished observations). Fortunately, serious complications after adenotonsillectomy are infrequent. Postoperative hemorrhage, however, does occur and can become a life-threatening catastrophe (discussed later). Postoperative emesis is relatively common because of pharyngeal mucosal irritation from surgery, which may stimulate the glossopharyngeal nerves, or from bloody secretions that are swallowed. It is therefore prudent for the anesthesiologist to administer antiemetics (e.g., ondansetron or metoclopramide) prophylactically, in addition to dexamethasone before the end of surgery (Pappas et al., 1998). Pain after tonsillectomy is mild to moderate and may be controlled by opioids given intraoperatively as part of anesthetic management or by supplementation (morphine, 50 mcg/kg; fentanyl, 0.5 to 1.0 mcg/kg; hydromorphone, 10 mcg/kg) in the postanesthetic care unit. Postoperative pain after adenoidectomy is relatively mild. Acetaminophen is a good adjunct to pain control in these patients.

It has been reported that children with OSAS have reduced opioid requirements after adenotonsillectomy. In addition, children with OSAS spontaneously breathing during halothane anesthesia are more likely to become apneic than non-OSAS children when treated with the same dosage of fentanyl (Waters et al., 2002). However, others have found no morbidity from opioid or benzodiazepine treatment in OSAS pediatric patients undergoing adenotonsillectomy (Helfaer et al., 1996; Sanders et al., 2006; Leong and Davis, 2007).

Preoperative sedation with benzodiazepines and postoperative treatment with opioids should be tailored to meet the needs of the patient. It would be prudent to titrate opioids, anticipating that the more severe the OSAS, the higher is the risk of opioid and benzodiazepine sensitivity. For patients with moderate or severe OSAS, many anesthesiologists prefer not to administer opioids until the patient has fully emerged from the anesthetic, has regained spontaneous ventilation, has been extubated, and is able to maintain an open airway with spontaneous ventilation. The patient’s opioid therapy may then be titrated carefully to achieve analgesia without respiratory compromise (Leong and Davis, 2007). During the immediate postoperative period, monitoring these patients in a postanesthetic care unit staffed with pediatric recovery room nurses allows a safe postoperative recovery with pain control.

Children with OSAS have a higher incidence of postoperative respiratory complications, including prolonged oxygen requirements, airway obstruction requiring nasal airway, and major respiratory compromise requiring airway instrumentation, than children without OSAS (Biavati et al., 1997; Wilson and Robertson 2002). Because OSAS is a disorder of anatomic and dynamic factors of upper airway function, it is not surprising that a procedure that is curative in many of these children would cause significant postoperative respiratory complications. Because these children are presumed to have impaired neuromuscular control of upper airway patency, residual anesthetic effects combined with blood, edema, and residual lymphoid tissue obstructing the postsurgical airway can lead to postoperative respiratory events (Wilson et al., 2002). McColley and colleagues (1992) reported a 58% incidence of severe respiratory compromise in children younger than 3 years that was associated with severe oxygen desaturation (SpO₂, ≤ 70%) or hypoventilation from upper airway obstruction. Among children with OSAS undergoing adenotonsillectomy (mean age, 5 years), intraoperative and postoperative complications are particularly high among those who were born prematurely (85% versus 25% to 34% in full-term infants).

Nixon and coworkers (2005) demonstrated that, not infrequently, children with OSAS continue to have polysomnogram- documented airway obstruction the night after adenotonsillectomy. Obstructive apneas and hypopneas were four times as frequent and associated with greater oxygen desaturation in the severe OSAS group as opposed to the mild OSAS group.

The subpopulation of children with OSAS who must be monitored in the hospital is still unknown. Children who are most likely to experience postoperative respiratory complications and have a higher postoperative respiratory disturbance index on their postoperative polysomnogram include children 3 years of age and younger, children with severe OSAS diagnosed by preoperative polysomnography, and those with associated medical conditions such as hypotonia, morbid obesity, failure to thrive, or severe structural airway abnormalities (Statham et al., 2006; Karlson 2008). These children are not candidates for outpatient surgery facilities and should receive medical care in centers with pediatric inpatient facilities and pediatric intensive care support. High-risk patients should be monitored overnight with continuous pulse oximetry, because standard apnea monitoring is unable to detect obstructive apnea and hypopnea. Patients can be discharged when significant oxygen desaturation during sleep has resolved.

A rare complication of adenotonsillectomy in children with severe OSAS is pulmonary edema resulting from relief of upper airway obstruction (Mehta et al., 2006). The exact cause is unknown. Galvis and colleagues (1980) hypothesized that during obstructed breathing, extreme negative pressures occur in the intrapleural and intrathoracic compartments. Intubation results in sudden equalization of pressure. The pulmonary venous pressure is suddenly much higher than intrathoracic pressure, resulting in pulmonary hyperemia and edema. Furuhashi-Yanaha and colleagues (2000) hypothesized that the abrupt switch from nasal to oral breathing during emergence from anesthesia might create acute obstruction, which creates an increase in negative intrapleural pressure, leading to pulmonary edema. Regardless of the cause, these children require oxygen and possible reintubation to provide CPAP or positive end-expiratory pressure. This condition usually resolves in 24 to 48 hours (Carcillo, 2003), and the use of diuretics is controversial.

Bleeding after Tonsillectomy

Occasionally, bleeding continues or recurs after tonsillectomy, and the child must be anesthetized to suture or pack the bleeding area. Posttonsillectomy bleeding is classified as primary or secondary bleeding. Primary bleeding occurs within the first 24 hours after surgery and is usually more brisk and profuse than secondary bleeding. Most fatal hemorrhages occur within the first 24 hours after surgery (Randall and Hoffer, 1998). Most posttonsillectomy bleeding, however, occurs within the first 6 hours after surgery (Crysdale and Russel, 1986; Carithers et al., 1987; Guida and Mattucci, 1990).

Secondary bleeding usually occurs 24 hours to 5 to 10 days after surgery, when the eschar covering the tonsillar bed sloughs (Verghese and Hannallah, 2001). Allen and coworkers (1973) reported a 0.1% incidence of posttonsillectomy bleeding necessitating repeat exploration in the operating room. Crysdale and Russel (1986) reported a 2.15% incidence of bleeding during the overnight stay in the hospital for 9400 children undergoing tonsillectomy with or without adenoidectomy. Of these, 3% (0.06% of all patients) required reexploration under general anesthesia.

The child with tonsillar bleeding may be brought back to the anesthesia service several hours or even days or weeks after the surgery. Because primary bleeding tends to be more rigorous, bleeding may obstruct the view of the larynx, and emergent tracheostomy may be needed. The otolaryngologist should be present before anesthetic induction. Whether the bleeding is primary or secondary, several issues must be addressed before induction of anesthesia; the child usually is hypovolemic, anemic, agitated, or in shock, with a stomach full of blood clots and often without an intravenous catheter. This is one of the most dangerous and challenging situations in pediatric anesthesia practice, and the failure to initiate prompt action has been one of the main causes of death associated with tonsillectomy (Alexander et al., 1965).

Under these circumstances, intravenous access must be reestablished, unless it is still in place, to rehydrate or transfuse without delay. It is frequently difficult to find a suitable vein for insertion of a large-enough catheter (at least 22 gauge for infants, 20 gauge for children) for rapid hydration and transfusion in an agitated and hypovolemic child with extreme cutaneous vasoconstriction. An intraosseous infusion or surgical cutdown may be indicated in some patients.

It may be difficult to estimate the extent of blood loss and dehydration. The patient may be tachycardic and have an elevated blood pressure caused by the release of endogenous catecholamines from hemorrhage, hypovolemia, or fear and excitement. If the child is sitting up and talking without feeling dizzy, hypovolemia is only mild to moderate; there is enough time to evaluate volume status and to determine hemoglobin, hematocrit, and urine specific gravity and obtain additional information on overall physical status. The child who is lying down and has pale conjunctivae, hypotension, and diminished consciousness may be on the verge of hypovolemic shock, and volume resuscitation must begin without delay. In most cases, tonsillar bleeding is not massive, and the child should be rehydrated and transfused as needed before the induction of anesthesia. Hematocrit values should be rechecked after rehydration to have a more accurate estimate of blood loss.

The anesthesiologist must be aware that the child and parents are extremely frightened. If the condition permits, the patient may be sedated with intravenous midazolam to relieve anxiety. Before the induction of anesthesia, precautions for a patient with a full stomach should be taken. These include at least one, but preferably two, well-functioning suction apparatuses with large-bore suction tubes, extra laryngoscope handles and blades, and several cuffed endotracheal tubes with lubricated stylets in place. An experienced assistant should always be available to help the anesthesiologist and to apply cricoid pressure during the rapid-sequence induction (see Chapter 13, Induction, Maintenance, and Recovery).

After preoxygenation, atropine (0.02 mg/kg), and a defasciculating dosage of nondepolarizing muscle relaxant in older children, the child is intubated, most commonly with a rapid-sequence technique using propofol (2 mg/kg), sodium thiopental (3 to 6 mg/kg), etomidate (0.3 to 0.4 mg/kg), or ketamine (2 mg/kg), with dosages guided by the patient’s hemodynamic status, and succinylcholine (2 mg/kg) with cricoid pressure. Alternatively, rocuronium (1.2 mg/kg) may be used. When using thiopental and rocuronium for induction, significant precipitation occurs if they are given together without flushing the intravenous tubing between the two drugs. If the child is still anemic or dehydrated before induction, ketamine (2 mg/kg) may be used instead of propofol or thiopental, because the latter two drugs may cause profound myocardial depression and hypotension. A cuffed endotracheal tube is used in these cases to prevent blood from entering the trachea around the uncuffed tube.

Management of anesthesia during maintenance and emergence in these patients focuses on hypovolemia and a full stomach. A large-bore gastric tube is introduced through the mouth to decompress the stomach, although it is not possible to evacuate all blood clots and solid food. Both sides of the chest are carefully auscultated, and the endotracheal tube is suctioned to rule out aspiration of blood or gastric contents. When there is doubt, fiberoptic bronchoscopy may be performed before extubation, and a portable chest radiograph is taken in the operating room or in the postanesthetic care unit. The endotracheal tube is left in place until the child is fully awake and normal gag and cough reflexes have returned.

The death of a child as a consequence of hemorrhage after tonsil and adenoid surgery is particularly tragic because the operation in most cases is elective and the child is in relatively good health. Mortality rates reported in the literature range from 1 in 1000 to 1 in 27,000 (10 to 0.37 per 10,000) in old reports (Bluestone et al., 1975; Avery and Harris, 1976). The anesthesia-related mortality unadjusted for age is reported to be 1 in 14,000 (Avery and Harris, 1976). A more recent survey in Europe shows the posttonsillectomy mortality rates of 0.62 to 0.63 per 10,000 (Windfuhr et al., 2009). Windfuhr and colleagues (2008a, 2008b) also reported life-threatening posttonsillectomy hemorrhage in 79 patients between 1980 and 2006. There were 36 children involved in the study. Out of 29 deaths, 18 were children. Nearly 90% of the life-threatening crises were related to secondary hemorrhage (>24 hours), and all deaths were directly or indirectly attributable to posttonsillectomy hemorrhage.

Peritonsillar Abscess

Peritonsillar cellulitis and abscesses occur more frequently in older children and adults than in young children. Most infections appear to originate in the tonsils and spread to the peritonsillar space between the tonsillar capsule and the superior constrictor muscle. The infection may spread upward into the palate, but it usually does not invade the posterior tonsillar pillar or the posterior pharyngeal wall (Teele, 1983). The patient presents with complaints of severe sore throat, difficulty in swallowing, and high fever. Progressive difficulty in opening the mouth may develop because of spasm of the pterygoid muscles. It is therefore imperative to determine the extent of limitation of mouth opening preoperatively. The affected tonsillar area is markedly inflamed and swollen, and the uvula is displaced to the opposite side (Cook, 1982). In older children and adolescents, it is possible to perform incision and drainage of a peritonsillar abscess with intravenous sedation using an opioid and topical or local anesthesia, with the patient in a head-down position and the head turned to the side of the abscess. If this is not possible, general endotracheal anesthesia is necessary.

The child should be well sedated. Anesthesia is induced with oxygen, nitrous oxide, and sevoflurane by mask or with intravenous sodium thiopental or propofol. Anesthesia is deepened in oxygen without nitrous oxide, while spontaneous breathing is maintained and assisted with a head-down tilt and with the head turned toward the affected side. Under deep inhalational anesthesia, the larynx is exposed carefully with the laryngoscope blade for intubation with a cuffed endotracheal tube. It is advantageous to let the patient breathe spontaneously to maintain the upper airway muscle tone and airway patency. Muscle relaxants should not be given until the anesthesiologist is certain that the airway can be maintained by mask and bag. Two large-bore suction catheters and tonsil suction tips should be on hand in case the abscess ruptures during the process of laryngoscopy and intubation.

Use of the Laryngeal Mask Airway During Adenotonsillectomy

Since the early 1990s, several reports (but no scientific studies) have described the use of the armored LMA rather than an endotracheal tube for airway support during tonsil or adenoid surgery. The armored LMA can be held within the Boyle-Davis gag and remain patent (Alexander, 1990) (Fig. 24-6). A reported benefit of using the LMA is avoiding tracheal intubation and its associated complications (trauma, cardiovascular stimulation, endobronchial intubation, coughing, laryngospasm, and subglottic edema) (Hatcher and Stack, 1999). The disadvantages of the armored LMA include the risk for aspiration of stomach contents and the risk of inadequate positioning (Hatcher and Stack, 1999).

FIGURE 24-6 Armored laryngeal mask airway used for adenotonsillectomy in a child.

(From Kretz FJ, Reimann B, Stelzner J, et al: The laryngeal mask in pediatric adenotonsillectomy: a meta-analysis of medical studies [in German]. Anaesthesist 49:706, 2000. With permission of Springer Science and Business Media.)

Hern and coworkers (1999) reported the surgeon’s perspective on the use of LMA for tonsillectomy. In their study, 44 children were anesthetized using an LMA, and 47 patients were anesthetized using an endotracheal tube before tonsillectomy. There was an 11.4% failure rate for the LMA, requiring change to an endotracheal tube before the end of the procedure. The LMA was reported to hamper the surgeon’s visualization of the field and might have led to removal of less tonsillar tissue in the LMA group. It is difficult to suggest that the LMA is a superior choice for airway management during tonsillectomy. Additional objective data must be accumulated before recommending this technique.

Risks Associated with Laser Surgery

Because of the potential danger with the use of laser beams in the operating room, the users and other operating room personnel should follow the safety guidelines established by the American National Standard for the Safe Use of Laser in Health Care Facilities and published by the American National Standard Institute (the ANSI Z136 Standard) (ANSI, New York). Doors to operating rooms where laser surgery is in progress should be clearly marked to prevent unprotected personnel from accidentally straying into the path of laser beams. Warning signs should specify the class and type of the laser being used as well as appropriate protective measures specific for the wavelength (Rampil, 1992; Pashayan, 1994).

All laser beams for medical use are transmitted through air and are well reflected by smooth metal surfaces. Patients and health care personnel involved in laser surgery are at risk for laser injury, particularly of the eyes. The nature and extent of eye injury depend on the wavelength of the laser beam and its energy. Because the laser beam of far-infrared wavelength is absorbed by the surface it first encounters, errant CO₂ laser beams can cause serious corneal ulcerations and scar formation (Liebowitz and Peacock, 1980). Visible and near-infrared lasers (e.g., KTP, argon, Nd:YAG) penetrate through the transparent cornea and anterior chamber of the eye, are absorbed by the retina, and cause serious damage. The anesthetized patient’s eyes should be taped closed and covered with saline-soaked eye pads or metal shields, or both (Keon, 1988). Operating room personnel must wear safety goggles that are specific for the laser wavelength in use. Safety goggles should provide wraparound protection from reflected beams (Rampil, 1992). For CO₂ lasers, any plastic or regular eyeglasses (not contact lenses or thin plastic splash protection shields) will suffice as protection (provided that lateral aspects of the eye are also covered), because far-infrared beams are completely absorbed. Visible and near-infrared lasers require specific filters for particular laser wavelengths.

Incandescent debris resulting from the explosive vaporization of the tissue in contact with the laser beam produces a plume of smoke with fine particulates (0.1 to 0.8 mcm), which may contain infectious or mutagenic viral particles (Nezhat et al., 1987; Kokosa and Eugene, 1989). Fragments of DNA have been found in the smoke plume, but only one study has found viral transmission from the laser plume (Garden et al., 2002). The same risk exists for the smoke plume from electrocautery. Ordinary surgical masks do not effectively filter particles smaller than 3.0 mcm; special high-efficiency masks are therefore needed to filter laser plume particles (Rampil, 1992).

A laser beam can ignite inflammable materials used for general anesthesia, such as endotracheal tubes, anesthesia circuits, oil-based lubricants, ointments, sponges, and drapes (Snow et al., 1976; Cozine et al., 1981; Hermens et al., 1983). All materials used for endotracheal tubes, with the exception of special metallic tubes, are quite vulnerable to laser impact (Geffin et al., 1986; Sosis, 1989a, 1989b). Polyvinyl chloride (PVC) tubes appear to be more easily punctured and ignited by CO₂ laser than red-rubber tubes, and to produce more toxic combustion fumes (Ossof et al., 1983), although black smoke from burning rubber is probably just as damaging to the airway mucosa. Wolf and Simpson (1987) found that once PVC is ignited, it is less flammable than silicone or red rubber, with a flammability index (i.e., the fraction of inspired oxygen [Fio₂] at which the material ignites) of 0.26, compared with silicone (0.19) and red rubber (0.18); this means that silicone and rubber tubes continue to burn in room air, whereas PVC tubes do not. Red-rubber tubes, however, are more resistant to puncture than PVC tubes (41 versus 0.8 seconds with PVC, at the same power density) and intramural fires are less likely to occur (Ossof, 1989). For these reasons, metal tubes are preferred overall.

During laser surgery, when a nonmetallic endotracheal tube is used, the Fio₂ should be reduced to 30% or less, as long as the patient is adequately oxygenated, to reduce combustibility of endotracheal tubes (Hermens et al., 1983). Oxygen may be diluted by nitrogen or helium to reduce the flammability of endotracheal tubes (Pashayan and Gravenstein, 1985; Simpson et al., 1990). Oxygen concentrations up to 60% may be used if helium is the diluent gas. Nitrous oxide supports combustion above 450° C and should therefore be avoided (Wolf and Simpson, 1987; Keon, 1988). It is imperative that the patient be immobilized during laser surgery. The choice of anesthetics depends largely on the technique of ventilation during the laser surgery.

Ventilatory Management During Laser Surgery

Numerous anesthetic approaches have been reported for airway management during laryngeal laser surgery. General transtracheal anesthesia, through a preexisting tracheostomy, is the easiest and probably the safest method. It is performed by replacing the preexisting tracheostomy cannula with a metal tracheostomy tube. Manually operated intermittent jet ventilation (Rontal et al., 1980; Scamman and McCabe, 1986; Ravussin et al., 1987) or high-frequency jet ventilation (Smith et al., 1975) by means of a needle or a catheter through the cricothyroid membrane has been used. These techniques, however, have a potential risk for barotrauma to the upper airways and overdistention of lungs (e.g., pneumothorax, pneumomediastinum), particularly when the laryngeal airway is obstructed. Subcutaneous emphysema and pneumomediastinum also may occur because of needle misplacement or a direct leak at the tracheal puncture site (Borland and Reilly, 1987). Without tracheostomy or transtracheal puncture, the maintenance of airways and pulmonary ventilation during laser surgery can be accomplished by means of manual jet ventilation (i.e., Saunder’s jet) without an endotracheal tube. A combination of intravenous propofol, topical lidocaine, and small dosages of opioids (i.e., fentanyl or remifentanil infusion) has been used successfully with the patient breathing spontaneously (Borland LM, unpublished data).

Nonflammable Endotracheal Tubes

Red-rubber tubes wrapped with reflective aluminum or copper adhesive-backed tape, originally described by Snow and colleagues (1974), have been used effectively (Norton et al., 1976), but the use of these homemade, laser-resistive tubes has decreased considerably over the past decade with the advent of commercially available nonflammable endotracheal tubes. Before wrapping a red-rubber tube with a metal tape, the tube should be wiped with alcohol to remove greasy or oily residues that interfere with adhesion. It should also be wiped with tincture of benzoin or an equivalent before spiral wrapping with a metal tape (Rampil, 1992). There should be no windows of exposed tube; care should be taken to prevent wrinkles, which may cause abrasion of the laryngotracheal mucosa.

Certain metallic tapes do not sufficiently prevent the CO₂, Nd:YAG, or KTP laser beams from penetrating the endotracheal tube and igniting a blowtorch fire. The CO₂ laser can ignite the adhesive backing of all these tapes and perforate the aluminum tapes within 0.1 second (Sosis, 1989a). Other aluminum tapes (3M No. 425 and 433; 3M Corp., St. Paul, MN) and a copper foil tape (Venture Tape Corp., Rockland, MA; 3M Corp.) are effective in shielding at least 60 seconds of direct exposure (Sosis and Dillon, 1990). A metallic foil wrap for endotracheal tubes, specifically manufactured for the laser, is also available (Laser Guard, Merocel Corp., Mystic, CT) and is approved by the U.S. Food and Drug Administration. This metal-based surgical sponge provides protection against CO₂, argon, and KTP lasers but not against Nd:YAG lasers, according to the study by the Emergency Care Research Institute (ECRI, 1990). It also provides a smoother surface than metal tapes on the endotracheal tube for better airway mucosal protection, but it is bulky, adding about 2 mm to the diameter of the endotracheal tube.

A number of laser-resistant endotracheal tubes with metal exteriors have become available for clinical use. The Laser Flex tube (Mallinckrodt, St. Louis, MO; Bivona, Gary, IN) is a stainless steel spiral with a PVC tip, with or without two distal saline-inflatable cuffs (Fig. 24-8). This tube is resistant to CO₂ and the Nd:YAG laser. The Fome Cuff tube (Bivona) is an aluminum spiral tube with an outer coating of silicone that is approved by the manufacturer for the CO₂ laser only. This tube has a self-inflating foam sponge–filled cuff. The foam in the cuff prevents deflation of the cuff after puncture, but damage to the filling tube may cause difficulty in deflating the cuff (Rampil, 1992). The Laser Shield II (Medtronic/Xomed, Jacksonville, FL) is silicon-based endotracheal tube wrapped with metal foil and then wrapped in polytetrafluoroethylene tape to smooth the tube. These tubes are available only in two sizes, and the smaller size (5.0) has an outside diameter with a cuff roughly equivalent to a size 5.5 endotracheal tube.

FIGURE 24-8 Flexible stainless steel (Laser Flex) tubes. Top to bottom: A 3.5-mm-ID, uncuffed tube; a 4.0-mm-ID, uncuffed tube; and a 5.0-mm, double-cuffed tube. ID, inner diameter.

(Courtesy Mallinckrodt Medical, Inc., St Louis, MO.)

Anesthesia with Manual Jet Ventilation

A more commonly accepted and practiced method for ventilation during laryngeal laser surgery is the use of a Venturi apparatus (i.e., Saunders’ jet injector), originally described by Saunders (1967) and Spoerel and coworkers (Donlon and Benumof (1996) and applied in pediatric patients by Miyasaka and colleagues (1980) as well as others (Johans and Reichert, 1984; Scamman and McCabe, 1986; Borland and Reilly, 1987). The apparatus consists of an injector and tubing, a toggle switch, a reducing valve or regulator, and tubing with a connector to a high-pressure wall oxygen supply (50 psi) (Figs. 24-9 and 24-10). After a precordial stethoscope, a pulse oximeter sensor, electrocardiogram leads, and a blood pressure cuff are positioned on the child, and anesthesia is induced, usually by mask with nitrous oxide, sevoflurane, and oxygen. An intravenous catheter is inserted, an axillary or rectal temperature probe is placed, and a nerve stimulator is positioned. Cisatracurium (or other intermediate-acting muscle relaxant), fentanyl (2 to 3 mcg/kg), and lidocaine (1 mg/kg) are injected intravenously. Midazolam may be given to ensure amnesia during the laser surgery. Alternatively, anesthesia may be switched to intravenous propofol with or without a small dosage of fentanyl. Nitrous oxide is discontinued, and the patient is ventilated with 100% oxygen for several minutes before the mask is removed. The surgeon positions the child’s head and inserts a Jako suspension laryngoscope to secure a patent laryngeal airway.

FIGURE 24-9 The Venturi ventilation apparatus with injector needle-on-clamp. A, High-pressure gas source with optional oxygen–nitrous oxide or helium blender. B, High-pressure tubing. C, Reducing valve and pressure gauge. D, Short piece of high-pressure tubing. E, Manual, all-or-none trigger valve. F, Tapered metal connector of the pop-off type. G, Low-pressure plastic tubing. H, Optional Luer-Loc fitting on tapered metal connector of the pop-off type. I, An adapted, curved stainless steel needle, 12 to 18 gauge. J, Screw clamp to fit on the edge of the operating laryngoscope. K, The operating laryngoscope. L, Laryngoscope handle that fits the suspension apparatus. It should be possible for fittings F and H to pop off when high pressure is inadvertently transmitted through the low-pressure tubing (i.e., when C or E, or both, are failing) or when an obstruction occurs distally.

(Adapted from Van Der Spek AF, Spargo PM, Norton ML: The physics of lasers and implications for their use during airway surgery, Br J Anaesth 60:709, 1988.)

FIGURE 24-10 A Saunders-type, hand-held jet ventilation apparatus.

(Courtesy Manujet III, VBM Medizintechnik, Sulz am Neckar, Germany.)

In children, supraglottic jet ventilation is used. Subglottic jet ventilation is rarely used in children, because it requires extremely delicate skills to access adequate egress of air between pulses, and if this is not accomplished, pneumothorax or pneumomediastinum results (Hunsaker, 1994).

For supraglottic ventilation, a 10-French metal injector is inserted into the side channel of the suspension laryngoscope and secured with the distal end 1 to 2 cm above the glottic opening; the metal injector is readjusted as needed to obtain the optimal distance and direction. Oxygen concentration of the jet injector is adjusted with an oxygen blender to reduce the inspired oxygen concentration with room air entrainment to 30% or less, provided that oxygen saturation on the pulse oximeter is 98% or higher. The anesthesiologist begins jet ventilation by intermittently squeezing the toggle switch. Short bursts of inspiration with air entrainment (1 to 2 seconds) with a driving pressure of 15 psi or less (the pressure may vary with different setups), followed by enough expiration time (2 to 4 seconds) to complete exhalation, provides adequate pulmonary ventilation. Close observation of the child’s chest motion during each breath is crucial to prevent air trapping, hyperinflation, and volutrauma to the lungs (Borland and Reilly, 1987). Inspired oxygen concentration is decreased as low as possible by adjusting the oxygen-blender to maintain SpO₂ between 98% and 100% throughout the procedure. If passive exhalation is unusually prolonged, measures to improve airway patency must be taken immediately. This can usually be accomplished by readjusting or redirecting the injector. Occasionally, when papillomatous growth about the glottis is extensive, the child may require intubation with a small tube and manual ventilation while the surgeon improves the glottic aperture by debulking papillomas on and around the vocal cords.

After the laryngeal surgery is completed, ventilation is maintained with 2:1 or 1:1 ratio of oxygen and air by mask and bag, or the airway is secured by intubation with a small cuffed or uncuffed orotracheal tube. Neuromuscular blockade is reversed in the usual manner. The endotracheal tube is removed after neuromuscular blockade is completely reversed and the child is breathing satisfactorily.

The manual jet ventilation technique offers several advantages in laryngeal laser surgery, including adequate ventilation, excellent surgical access to the larynx (especially to the anterior and posterior commissures where an endotracheal tube obliterates the exposure), and the absence of ignitable material in the surgical field (Norton et al., 1976; Goforth et al., 1983). Potential risks of this technique include pneumothorax, pneumomediastinum, distention of the stomach and regurgitation, aspiration of resected material, and dehydration of mucosal surface (Lines, 1973; Norton et al., 1976; Rontal et al., 1980; Ruder et al., 1981; Keon, 1988).

Laser Bronchoscopy

The laser may be used coupled with a ventilating bronchoscope. If the CO₂ laser is used with the rigid ventilating laser bronchoscope, higher oxygen concentrations can be employed, because there is no flammable material in the airway. The fibers used for KTP and Nd:YAG laser delivery may support combustion, and when used through a bronchoscope, the inspired oxygen concentration should be less than 40%. The newer CO₂ laser waveguide has not been tested for combustibility, so prudence suggests limiting the inspired oxygen concentration to less than 40%.

Anesthetic Approach for Tracheostomy

The need for tracheotomy may occur emergently, urgently, or electively. It is essential for the surgeon and anesthesiologist to establish a plan before surgery. Some situations, including laryngeal trauma, massive facial injuries, upper airway obstruction, and oropharyngeal distortion, require emergency tracheostomy for airway management. This procedure is usually performed outside the operating room and, fortunately, rarely in the pediatric population.

Urgent tracheotomy may be required in neonates for congenital airway defects and for older children because of upper airway trauma, corrosive ingestion, infections, and tumors (see Table 24-6). Although these children may initially maintain oxygenation and ventilation, they are at risk for acute respiratory failure. They may have airway anomalies that contribute to difficult or impossible tracheal intubation. The location of the obstruction dictates the surgical and anesthetic approach. The anesthesiologist must establish whether the child can maintain a patent airway under anesthesia and whether they can be intubated by standard laryngoscopy or fiberoptic bronchoscopy. Sedation with ketamine may also be an option. Otherwise, anesthesia is induced with an inhaled anesthetic agent after preoxygenation in the position in which the child is most comfortable, often the sitting position. If no intravenous catheter is in place, one is inserted, and atropine (0.01 to 0.02 mg/kg) or glycopyrrolate (0.01 mg/kg) may be injected intravenously as indicated after induction of anesthesia. Unless a normal, unobstructed airway can be maintained, muscle relaxants should not be used because total upper airway obstruction may result. After the child is anesthetized but spontaneously breathing, topical lidocaine (up to 5 mg/kg) may be sprayed on the larynx to ablate laryngeal reflexes before laryngoscopy. If intubation is impossible, an LMA may be used to maintain an airway until the tracheotomy is performed. Tracheostomy is best performed in infants with an endotracheal tube or a rigid bronchoscope in place to ensure adequate pulmonary gas exchange. The endotracheal tube also helps the surgeon to identify the trachea. Many surgeons prefer to use a rigid bronchoscope held in place rather than an endotracheal tube. The rigid bronchoscope allows a firmer anatomic reference for the trachea, but it cannot be taped in place and must be held by hand. Elective tracheotomy is usually performed in children already intubated for chronic ventilation or pulmonary toilet or performed in combination with tumor removal or repair of tracheal abnormalities.

For tracheotomy, the patient is positioned with the shoulders elevated by placing a roll under the neck so that the neck is hyperextended. Some surgeons request that the anesthesiologist hold the chin with the left hand to keep the neck stretched so that soft tissue over the trachea is stabilized (Fig. 24-11) (Stool and Eavey, 1990). The surgeon palpates the neck to identify the thyroid and cricoid cartilages, sometimes with difficulty. The trachea is palpated down to the sternal notch, guided by the presence of an endotracheal tube. A skin incision is made about one fingerbreadth above the sternal notch, either horizontally or vertically. When the fascial layer is incised under the subcutaneous fat and the trachea is identified, a pair of stay sutures, one on each side of the tracheal incision, is placed for traction during the cannulation of the tracheotomy tube. The stay sutures also serve to prevent asphyxiation in the event that the tracheotomy tube becomes displaced in the immediate postoperative period before the tract has formed (Myers et al., 1985). Before the tracheotomy cannula is inserted in the tracheal incision, the endotracheal tube or bronchoscope is slowly pulled back to accommodate the insertion. The endotracheal tube or bronchoscope should not be removed completely until successful ventilation is demonstrated after placement of the tracheotomy tube. The anesthesia delivery tube is then disconnected from the endotracheal tube under the drapes and reconnected to the proximal universal adapter end of the tracheotomy tube over the operative field.

FIGURE 24-11 A child positioned for tracheostomy with the shoulders elevated, the neck hyperextended, and the chin pulled up by hand.

An esophageal stethoscope or a nasogastric tube has occasionally been misidentified as an endotracheal tube in the trachea by the surgeon, and inadvertent esophageal incision has been made (Schwartz and Downes, 1977); esophageal stethoscopes therefore should be avoided during tracheotomy. The child is monitored with a precordial stethoscope, pulse oximeter, capnography, electrocardiography, automated blood pressure cuff, and rectal or skin thermometer.

Major intraoperative and postoperative complications of tracheotomy include hemorrhage, subcutaneous emphysema, pneumothorax, and pneumomediastinum. The neck and shoulders of the child should be palpated for the typical crepitus to rule out subcutaneous emphysema. Then both sides of the chest are auscultated. In most hospitals, the patient is transferred to the ICU for observation and a chest radiograph to ensure proper tracheotomy tube positioning and the absence of pneumothorax.

Subglottic Stenosis

Neonatal subglottic stenosis can be categorized as congenital or acquired. The incidence of congenital subglottic stenosis is 5%; the remaining cases are acquired. Subglottic stenosis in the full-term infant is defined as a subglottic airway diameter of less than 4 mm at the level of the cricoid cartilage, and less than 3 mm in the premature infant (Walner et al., 2001). This narrowing can be compared with the reference point of a neonatal endotracheal tube. The outer diameter of a 2.5 endotracheal tube is 3.6 mm, and the outer diameter of a 3.0 endotracheal tube is 4.2 mm (Rutter et al., 2003). The diagnosis of subglottic stenosis is made by rigid endoscopy while the patient is anesthetized. The extent of the subglottic stenosis is determined by placement of an endotracheal tube that allows an airway leak between 10 and 25 cm H₂O (Table 24-8). This leak allows accurate measure of the airway size. The airway stenosis is then graded as follows: grade I, less than 50% obstruction; grade II, 51% to 70% obstruction; grade III, 71% to 91% obstruction; and grade IV, no detectable lumen (Gerber and Holinger, 2003).

In the absence of trauma, an abnormality of the cartilage or subglottic tissues is usually considered congenital. The cause of congenital subglottic stenosis is thought to be a failure of the laryngeal lumen to recanalize after completion of normal epithelial fusion at the end of the third month of gestation (Walander, 1955). Congenital subglottic stenosis lies on a continuum of embryologic failure that includes laryngeal atresia, stenosis, and webs. In its mildest form, congenital subglottic stenosis merely represents a normal-appearing cricoid with a smaller-than-average diameter, usually with an elliptical shape. Infants and children with mild subglottic stenosis may present with a history of recurrent upper respiratory infections, often diagnosed as croup, in which minimal glottic swelling precipitates airway obstruction. The location of the stenosis is usually 2 to 3 mm below the true vocal cords (Rutter et al., 2003).

Severe congenital subglottic stenosis can be a life-threatening airway emergency manifesting immediately after the infant is delivered. See related video online at www.expertconsult.com.If endotracheal intubation is successful, the patient may require intervention before extubation. In more severe cases, tracheotomy can be life saving at the time of delivery. Infants with congenital subglottic stenosis may have surprisingly few symptoms and may not present for treatment for weeks or months after birth (Rutter et al., 2003). After the initial management of congenital subglottic stenosis, the larynx will grow with the patient and may not require further surgical intervention.

Laryngomalacia is a congenital condition of excessive flaccidity of the laryngeal structures, especially the epiglottis and arytenoids, most likely caused by the lack of neural control of laryngeal muscles. See related video online at www.expertconsult.com.Laryngomalacia accounts for more than 70% of persistent stridor in neonates and young infants (Fig. 24-12); the remaining 10% of neonatal stridor is caused by vocal cord paralysis (Yellon et al., 2002) (Fig. 24-13).

FIGURE 24-12 Laryngomalacia. A, The larynx during expiration. The omega-shaped epiglottis and the arytenoid cartilages appear normal. B, The larynx during inspiration. The force of the inspiratory airflow leads to collapse of the laryngeal inlet. Infolding of the epiglottic surface and the arytenoids causes severe upper airway obstruction.

(From Yellon RF, McBride TP, et al: Otolaryngology. In Zitelli BJ, Davis HW, editors: Atlas of pediatric physical diagnosis, ed 4, St Louis, 2002, Mosby, p 863.)

FIGURE 24-13 A, Congenital bilateral vocal cord paralysis. The marked narrowing of the aperture between the vocal cords stems from the loss of ability to abduct on inspiration. B, Normal neonatal vocal cords with a wide opening during inspiration.

(From Yellon RF, McBride TP, Davis HW: Otolaryngology. In Zitelli BJ, Davis HW, editors: Atlas of pediatric physical diagnosis, ed 4, St Louis, 2002, Mosby, p 863.)

In the mid-1960s, long-term intubation was introduced as the preferred artificial airway in neonates (McDonald and Stocks, 1965). As neonatal survival rates increased with the advent of neonatology and improved neonatal intensive care, subglottic stenosis occurred as a frequent morbidity caused by long-term tracheal intubation. Acquired subglottic stenosis results from intubation trauma and accompanying inflammatory responses. Factors suspected of contributing to this problem include prematurity, the size and amount of movement of the endotracheal tube, duration of intubation, laryngeal or tracheal injury during intubation, and presence of infection during the course of tracheal intubation. The incidence of acquired subglottic stenosis has been decreased over the past decade by minimizing the issues known to exacerbate subglottic injury (Walner et al., 2001). Patients presenting to the otolaryngologist for diagnosis or treatment of subglottic stenosis are former preterm infants who were intubated in the neonatal ICU for extended periods and have a tracheotomy in place with an established diagnosis, or are examined on an outpatient basis for symptoms suggesting subglottic stenosis (Cotton, 2000).

Laryngotracheal Reconstruction in Children

Chronically ventilated infants who have undergone tracheotomy because of prolonged intubation and subglottic stenosis will most likely require laryngeal reconstructive surgery. Children with greater than 70% obstruction of their laryngeal lumen usually require this surgery to allow decannulation (Rutter et al., 2003).

Laryngotracheal reconstruction has become the standard of care for symptomatic subglottic stenosis in the pediatric age group. There are five stages: characterization of the stenosis, expansion of the tracheal lumen, stabilization of the framework, healing of the airway, and decannulation (Rothschild et al., 1995). The procedure has evolved to include a variety of techniques for expanding the laryngotracheal complex to provide a stable airway of sufficient size. These include, but are not limited to, anterior cartilage graft with the tracheotomy left in place without stent, long-term (several months) stenting with or without cartilage grafts, and short-term (4 to 6 weeks) stenting with anterior or posterior cartilage grafts (Cotton, 2000).

Anterior cartilage graft with a tracheotomy left in place without a stent is indicated primarily for isolated anterior subglottic stenosis with no or relatively mild posterior subglottic components. A variation of this procedure is to remove the tracheotomy at the time of surgery and perform a single-stage laryngotracheoplasty. Posterior division of the cricoid plate and the introduction of a cartilage graft between the cut ends are indicated particularly for children with persistent posterior glottic pathology or primarily posterior subglottic pathology (Cotton, 2000).

Single-stage laryngotracheal reconstruction (LTR) uses cartilage grafts to obtain stability of the reconstructed airway. Single-stage LTR may include an anterior cartilage graft, a posterior cartilage graft, or both, and reconstruction often includes a cartilage graft at the former stoma site. The grafts are supported temporarily by a full-length endotracheal tube fixed in position through the nasal route (Cotton, 2000). The optimal time for extubation has not been established definitively. In general, children remain intubated for 7 to 10 days for anterior cartilage grafts alone, and 12 to 14 days if a posterior and anterior graft is required (Cotton, 2000).

The patient may be anesthetized by the intravenous route or with inhalation anesthesia through a tracheotomy cannula. Standard monitoring includes pulse oximetry, precordial stethoscope, capnography, anesthetic gas monitoring, electrocardiography, blood pressure apparatus, peripheral nerve stimulator, and axillary or rectal temperature monitor. The patient is placed in the tracheotomy position with the shoulders elevated and the neck hyperextended. A tracheotomy tube is replaced with a sterile cuffed armored (anode) endotracheal tube through the tracheostomy stoma and is covered under an adhesive drape to minimize contamination of the surgical field.

Through a horizontal skin incision over the cricoid cartilage, laryngeal structures are exposed. A vertical incision is made through the lower thyroid cartilage, the cricoid cartilage, and the upper tracheal rings (Fig. 24-14, A). The rib cartilage graft, which had been taken from the anterior chest wall with the external perichondrium intact, is used for the anterior or posterior grafts (see Fig. 24-14, B). Auricular cartilage and septal cartilage have also been used as graft materials. Toward the conclusion of surgery, the armored tube is removed from the tracheotomy stoma and replaced with a cuffed nasotracheal tube, one size larger than is appropriate for the patient’s age and size, to stent the larynx. This intraoperative transition needs to be carefully coordinated between the anesthesiologist and the surgeon.

FIGURE 24-14 Schematic of laryngotracheoplasty for subglottic stenosis. A, Laryngotracheal incision with upper and lower extension, depending on the extent of stenosis, and a cartilage graft. B, Cartilage graft is shown with the perichondrium internalized. A graft is sutured in position.

(From Licameli GR, Healy GB: Surgery of the larynx and trachea. In Bluestone CD, Stool SE, editors: Atlas of pediatric otolaryngology, Philadelphia, 1995, Saunders, pp 597–632.)

One requirement of single-stage LTR is meticulous postoperative management of the patient’s condition in the ICU. The nasotracheal airway must be maintained securely during the time of extended stenting without accidental extubation. The postoperative management of patients with single-stage LTR is not standardized (Cotton, 2000). Some centers use sedation to prevent agitation and accidental self-extubation and to avoid pharmacologic paralysis (Rothschild et al., 1995). However, many children do not tolerate nasotracheal intubation and require prolonged sedation and neuromuscular blockade to ensure maintenance of an airway with minimal trauma during healing (Brandom, 1997; Yellon et al., 1997). If neuromuscular blockade is used, daily recovery of neuromuscular function and the avoidance of prolonged use of corticosteroids will be associated with less muscle weakness after extubation (Yellon et al., 1997). Prolonged sedation with benzodiazepines or opioids can lead to withdrawal syndrome, and close observation after weaning from these agents is paramount. The role of dexmedetomidine in pediatric patients for prolonged sedation in the ICU setting continues to be investigated (Chrysostomou et al., 2006, 2009; Bejian et al., 2009).

Acute Supraglottitis or Epiglottitis

Acute infectious epiglottitis or supraglottitis is a relatively uncommon but truly life-threatening disease of childhood. See related video online at www.expertconsult.com.Acute epiglottitis is an acute bacterial infection principally involving supraglottic structures, including the lingular surface of the epiglottis, the aryepiglottic folds, and the arytenoids, with little or no involvement of subglottic structures, including the laryngeal surface of the epiglottis. It is therefore more appropriate to call it supraglottitis. Until the late 1980s, before the vaccination became available, Haemophilus influenzae type b (Hib) was the causative organism in more than 75% of cases. Group A β-hemolytic streptococci account for most of the rare cases of epiglottitis (Gorelick and Baker, 1994). With the advent of effective Hib conjugate vaccines in 1988, the incidence of Hib disease declined dramatically in the United States (Adams et al., 1993; Hoekelman, 1994). The age-specific incidence of Hib disease among children younger than 5 years old decreased by 71%, and that of Hib meningitis by 82%, between 1985 and 1991 (Adams et al., 1993). Similarly, the incidence of acute supraglottitis, one of the most dreaded childhood emergencies, has declined by 84% (Gorelick and Baker, 1994).

Until the late 1980s, before the era of Hib conjugate vaccination, preschool children between the ages of 2 and 6 years had the highest incidence of supraglottitis, although it occurred in infants, older children, and occasionally adults (Schloss et al., 1979; Blackstock et al., 1987; Crysdale and Sendi, 1988). The highest number of cases is reported in the spring and fall, but sporadic cases are reported in other seasons as well. The condition usually begins with complaints of severe sore throat associated with dysphagia and a thick, muffled voice without a preceding history of runny nose. Rarely is there a croupy or barking cough, except in atypical cases in infants (Blackstock et al., 1987). The symptoms rapidly progress, and typical pathognomonic signs of dysphagia, dysphonia, dyspnea, and drooling (the 4 Ds) appear (Diaz, 1985).

In most children presenting with stridor in the emergency room, the most common symptoms found in those with acute supraglottitis are drooling, agitation, and absence of spontaneous cough (Mauro et al., 1988). The child looks toxic and usually has a high fever and tachycardia. The disease progresses very rapidly and may be fatal, with severe airway obstruction within 6 to 12 hours, unless immediate steps are taken to restore the patient’s upper airway patency. Classically, the child is sitting up, is dyspneic with the mouth open and drooling, and resists attempts to get him or her to lie down. There is forward chin thrust, slight cervical flexion, and forward flexion at the waist, together with tripod placement of the arms to support this posture. The inspiratory phase is slow, with stridor and retraction, whereas the expiratory phase is unobstructed. The inspiratory sound has been compared with that of a quacking duck.

As soon as acute supraglottitis is diagnosed in the emergency room, a pediatric anesthesiologist, an intensive care physician, and an otolaryngologist should be called. The child should never be left unattended by one of these physicians, because the disease can progress so rapidly that complete upper airway obstruction may ensue within minutes. The patient is kept in a sitting or tripod position with an oxygen mask in place and with pulse oximetry monitoring. The child is transferred to the operating room without delay and is accompanied by an anesthesiologist or an intensivist and a otolaryngologist.

A laryngoscope with several blades, endotracheal tubes with stylets, a bronchoscope, and a self-inflating bag with face mask and oxygen must accompany the patient for emergency intubation (Oh and Motoyama, 1977). Throat examination should not be attempted in the emergency room, because the risk for complete airway obstruction is great. If the child’s condition permits or when the diagnosis is questionable, the patient may be transferred to the pediatric ICU, where a radiograph of the lateral neck is taken with a portable x-ray machine (Butt et al., 1988). In the ICU, all the help and expertise needed for an emergency should be readily available. In most patients, a swollen epiglottis obstructing the air space (shadow) is seen on the lateral neck radiograph (Fig. 24-15, A). Negative radiographs, however, do not necessarily rule out supraglottitis.

FIGURE 24-15 Radiographs of the lateral area of the neck. A, Supraglottitis. The upper airway is obstructed by marked swelling of the epiglottis and other supraglottic structures. Notice the loss of normal curvature of the cervical spine. B, Laryngotracheitis. Arrows point to airway constriction below the vocal cords.

(Courtesy Dr. K. S. Oh, Department of Radiology, Children’s Hospital of Pittsburgh.)

Compared with acute supraglottitis, the onset of laryngotracheobronchitis (subglottic croup) is insidious. The child starts with the symptoms of upper respiratory infection, with rhinorrhea, cough, sore throat, and low-grade fever for several days before developing the symptoms of upper airway obstruction characterized by inspiratory stridor and barky or seal-like coughs (i.e., croup). The symptoms may last for a few days to more than a week, with varying degrees of severity that is worse at night in the supine position. In severe cases, endotracheal intubation under general anesthesia is required.

A lateral neck radiograph typically shows a narrowing of the airway shadow below the vocal cords (see Fig. 24-15, B). An anteroposterior radiograph of the neck may show a long area of airway narrowing resembling the steeple of a church (“steeple sign”) (Fig. 24-16).

FIGURE 24-16 Subglottic croup (i.e., laryngotracheobronchitis). A, Anteroposterior radiograph of the neck reveals a long area of the airway (shadow) narrowing and extending well below the normally narrowed area at the level of the vocal cords. This finding often is called the steeple sign. B, In this child, direct visualization using a bronchoscope revealed subglottic narrowing that was so severe that it would require endotracheal intubation or tracheostomy to establish an adequate airway.

(From Yellon RF, McBride TP, Davis HW: Otolaryngology. In Zitelli BJ, Davis HW, editors: Atlas of pediatric physical diagnosis, ed 4, St Louis, 2002, Mosby, p 860; and courtesy Sylvan Stool, MD, Children’s Hospital of Pittsburgh.)

Anesthetic Management of Acute Supraglottitis

In the operating room, anesthesia is induced with sevoflurane and oxygen while the child is in a sitting position and monitored with a precordial stethoscope, pulse oximeter, and electrocardiogram. The child in respiratory distress usually breathes continually, uninterrupted by breath-holding. A moderate continuous positive pressure (10 to 15 cm H₂O) must be maintained to minimize the inspiratory collapse of laryngeal airways by the Venturi effect, and ventilation must be assisted with moderate continuous positive airway pressure while avoiding inflating the stomach with excessive pressure. An intravenous access is established as soon as the child is sufficiently obtunded, and atropine (0.02 mg/kg) is given to block vagally mediated slowing of the heart; bradycardia in an atropinized child is a sign of severe hypoxia.

The induction may take much longer, even with a higher concentration of an anesthetic, to reach the state where relaxation and centrally fixed pupils indicate that the child is ready for intubation. A moderate CPAP should be applied to expand inflamed pharyngeal soft tissues and prevent the collapse of the airway, especially during the inspiratory phase of assisted ventilation. The use of muscle relaxants is contraindicated, even when the maintenance of the airway with mask and bag seems reasonable, because their use often results in the relaxation of pharyngeal muscles and complete obstruction of the laryngeal airway; frantic attempts to ventilate the child with high pressure cause gastric distention, regurgitation, and further asphyxia.

Endotracheal intubation is performed orally with a styletted tube one or two sizes smaller than usual. Visualization of the classic cherry-red epiglottis under direct laryngoscopy confirms the diagnosis (Fig. 24-17). However, the picture may be atypical or only part of the epiglottis may be inflamed or swollen. Inflammatory swelling of other supraglottic structures, including the uvula, arytenoids, aryepiglottic folds, and false vocal cords, also causes severe obstruction. To visualize the vocal cords for intubation, the epiglottis is lifted by the curved tip of a straight laryngoscope blade (e.g., Phillips 1). In children with severe obstruction and mucosal swelling, identification of supraglottic structures and glottic opening with a laryngoscope blade may be extremely difficult. Under these circumstances, forcible manual chest compression by the assistant may open up the expiratory passages momentarily (or produce a few bubbles of expired air). By aiming the tip of the endotracheal tube at that spot and advancing the tube with a gentle twisting motion, the physician may be successful in inserting the tube in the glottis (Smith, 1980). In case intubation proves to be impossible, a surgeon who can perform an emergency tracheotomy in the operating room with the surgical instruments and various tracheotomy tubes should be immediately available.

FIGURE 24-17 A bright red and swollen epiglottis in a patient with acute supraglottitis. It may retain its omega shape (A) or resemble a cherry (B).

(From Yellon RF, McBride TP, Davis HW: Otolaryngology. In Zitelli BJ, Davis HW, editors: Atlas of pediatric physical diagnosis, ed 4, St Louis, 2002, Mosby, p 858.)

After the endotracheal tube is in place, breath sounds are checked on both sides of the chest, and the tube is fixed with adhesive tape. The breath sounds are monitored carefully with the precordial stethoscope while continuous positive pressure (10 to 15 cm H₂O) is maintained, because pulmonary edema may occur after the relief of severe upper airway obstruction (Travis et al., 1977; Galvis et al., 1980), with an incidence of about 7% (Davis et al., 1981). Throat and blood cultures are taken during this period of stabilization but not before the intubation and restoration of adequate pulmonary gas exchange.

After the child’s condition is stabilized and adequate alveolar ventilation is reestablished, some anesthesiologists replace the orotracheal tube with a nasotracheal tube that has a leak, at or below 30 cm H₂O, for postoperative management in the ICU with minimal sedation of the patient (Oh and Motoyama, 1977). This is not a necessary step, and the physician should be confident that the switch could be made without losing the airway. Because acute epiglottitis has become an uncommon phenomenon, many anesthesiologists are happy with the oral intubation and leave the tube in place. The child can be deeply sedated for 24 hours to avoid premature extubation.

Since the mid 1970s, the management of epiglottitis with tracheotomy has been replaced by the less invasive long-term nasotracheal or orotracheal intubation as the standard treatment of choice. This has significantly reduced the duration of hospital stay for children with acute supraglottitis (Oh and Motoyama, 1977; Schloss et al., 1983). Accidental or self-extubation is a potentially disastrous complication. Careful taping of the tube and proper restraint of elbows and hands can prevent untimely extubation. Because most children are bacteremic, antimicrobial therapy, consisting of ampicillin with sulbactam, cefotaxime, or ceftriaxone, is started after blood and throat cultures are obtained. Any fluid deficits should be corrected parenterally. After emergence from anesthesia, the child often falls back to sleep for some time because of sleep deprivation and exhaustion. After this period, most patients require minimal or no sedation (Oh and Motoyama, 1977), although deep sedation has been used in some centers. Usually, it is possible to extubate these patients in 24 to 48 hours. A return of normal body temperature and increased leaks around the nasotracheal tube are major signs of recovery. The epiglottis may be examined under direct vision, with intravenous propofol to determine the time for extubation. Alternatively, a flexible fiberoptic bronchoscope may be used transnasally with intravenous sedation.